Landsat-based Estimation of Fractional Cover for several Salinas Valley Horticultural Crops

Abstract:

Green fractional canopy cover (Fc) can facilitate monitoring of crop development and evapotranspiration (ET) by farm managers and other agricultural stakeholders. The CropManage decision-support system (DSS) operated by UC Cooperative Extension, for instance, uses pre-specified empirically derived Fc timeseries to support the development of daily ET estimates and irrigation recommendations for several crop types. In this study, a multi-year field campaign was conducted on commercial farms in California's Salinas Valley to explore the relationship of Fc with Landsat NDVI for high-value specialty crops including broccoli, Brussels sprouts, cauliflower, celery, lettuce, pea, and strawberry. Approximately 2000 nadir-view digital photos were collected at 83 fieldsites during the 2016-2019 growing seasons with a boom-mounted multispectral SLR camera. Image segmentation was applied to derive the proportion of green vegetation within each photo, and the resulting data were reduced to a total of 617 Fc observations. Landsat 7-8 NDVI timeseries, derived from the USGS Surface Reflectance product, were collected for comparison to the field measurements. The data were accessed by an API from NASA’s Satellite Irrigation Management Support (SIMS) (Melton et al., 2012; Pereira et al., 2020), as hosted on Google Earth Engine. Linear interpolation was applied by SIMS to estimate NDVI for days between satellite overpasses. For the pooled dataset, a linear relationship was observed between Fc and NDVI (r²=0.89; RMSE=0.089). For crop-specific relationships, r² ranged from 0.79-0.96 with RMSE 0.065-0.101. The CropManage DSS can now access satellite data from SIMS to further inform development of irrigation recommendations and thereby support effective use of irrigation water resources. This capability is expected to be of particular benefit when crop phenology varies from baseline due to anomalous growing conditions,or when plant populations diverge from conventional practices. Satellite linkage may also help CropManage expand to serve crops and regions that lack ground-based canopy data.